← Extractions

Testing Randomness — Jakub Rodzik at wroclove.rb 2023

Jakub Rodzik's wroclove.rb 2023 talk walks through a dice-game story-developer 'Fredo' who learns to test non-determinism by progressing from conditional tests, to stubs, to dependency injection with a fake dice, to event logging. Closes with tradeoffs of each technique and the core message: don't test randomness — understand it, extract it, and control it.

Model

claude-opus-4-7

Ingestion

2fe730c4

Input tokens

371,454

fresh: 251,430
cached: 110,777
cache write: 9,247

Output tokens

11,782

Duration

184.0s

Roundtrips

Tool calls

Cost

$0.00

Nodes/edges extracted

21 / 43

Read set (nodes/edges)

135 / 4

Nodes (21)

update Testing Randomness talk

description Talk at wroclove.rb 2023. → Jakub Rodzik's wroclove.rb 2023 talk, framed as a bedtime story about developer 'Fredo' writing a two-player dice gam...

short_description Talk at wroclove.rb 2023. → Talk on testing non-deterministic code using stubs, dependency injection, and event logging.

update Jakub Rodzik person

description Conference speaker. → Ruby developer and conference speaker. At wroclove.rb 2023 presented 'Testing Randomness', using a story about develo...

short_description Conference speaker. → Ruby developer; wroclove.rb 2023 speaker on testing randomness.

create Test Driven Development concept

kind (empty) → concept

name (empty) → Test Driven Development

slug (empty) → test-driven-development

attrs (empty) → {"category" => "methodology"}

description (empty) → Discipline of writing a failing test first, then implementing the minimum code to make it pass, then refactoring. In ...

short_description (empty) → Practice of writing tests before production code and letting them drive design.

create Open-Closed Principle concept

kind (empty) → concept

name (empty) → Open-Closed Principle

slug (empty) → open-closed-principle

attrs (empty) → {"category" => "pattern"}

description (empty) → The 'O' in SOLID — software entities should be open for extension but closed for modification. In Rodzik's talk, 'Fre...

short_description (empty) → SOLID principle: software entities should be open for extension, closed for modification.

create Fake Object concept

kind (empty) → concept

name (empty) → Fake Object

slug (empty) → fake-object

attrs (empty) → {"category" => "pattern"}

description (empty) → A test double implemented as a full class (not just a stub on a method) that mimics the interface of a production col...

short_description (empty) → Test double implemented as a real class with controllable behavior replacing a collaborator.

create Event Logging for Testability concept

kind (empty) → concept

name (empty) → Event Logging for Testability

slug (empty) → event-logging-for-testability

attrs (empty) → {"category" => "practice"}

description (empty) → Technique of appending an entry to an events/audit log on every state change (dice rolls, winner assignment, etc.) so...

short_description (empty) → Recording state-change events so tests and operators can assert on what actually happened.

create Logic in Tests concept

kind (empty) → concept

name (empty) → Logic in Tests

slug (empty) → logic-in-tests

attrs (empty) → {"category" => "practice"}

description (empty) → Anti-pattern covered in Rodzik's talk. Tests that contain conditionals (`if p1_score > p2_score then expect ping else...

short_description (empty) → Anti-pattern of putting conditionals, loops, or production helpers inside test code.

create Time.current Testing Trap concept

kind (empty) → concept

name (empty) → Time.current Testing Trap

slug (empty) → time-current-testing-trap

attrs (empty) → {"category" => "practice"}

description (empty) → Pattern Rodzik singles out at the end of the talk: `Time.current` (or `Time.now`) looks simple but is functionally a ...

short_description (empty) → Using Time.current/Time.now as a hidden source of non-determinism in tests.

create RBS tool

kind (empty) → tool

name (empty) → RBS

slug (empty) → rbs

attrs (empty) → {"category" => "tool"}

description (empty) → Ruby's official type-signature system (`.rbs` files). Suggested by an audience member in the Testing Randomness Q&A a...

short_description (empty) → Ruby's official type-signature language.

create Datadog tool

kind (empty) → tool

name (empty) → Datadog

slug (empty) → datadog

attrs (empty) → {"category" => "service"}

description (empty) → Observability SaaS mentioned in Rodzik's talk as one place where events/state-change logs can be stored — with the ca...

short_description (empty) → SaaS observability platform for metrics, logs and traces.

create Sumo Logic tool

kind (empty) → tool

name (empty) → Sumo Logic

slug (empty) → sumo-logic

attrs (empty) → {"category" => "service"}

description (empty) → Log-management SaaS mentioned in Rodzik's talk as a destination for event logs, with a warning that extensive logging...

short_description (empty) → SaaS log management and analytics platform.

create Practical Object-Oriented Design in Ruby resource

kind (empty) → resource

name (empty) → Practical Object-Oriented Design in Ruby

slug (empty) → practical-object-oriented-design-in-ruby

attrs (empty) → {"type" => "book"}

description (empty) → Sandi Metz's book on object-oriented design in Ruby, commonly abbreviated POODR. Rodzik references it (alongside 99 B...

short_description (empty) → Sandi Metz's foundational book on OO design in Ruby (POODR).

create 99 Bottles of OOP resource

kind (empty) → resource

name (empty) → 99 Bottles of OOP

slug (empty) → 99-bottles-of-oop

attrs (empty) → {"type" => "book"}

description (empty) → Book by Sandi Metz and Katrina Owen on incremental object-oriented design and refactoring. Rodzik cites it as a sourc...

short_description (empty) → Sandi Metz and Katrina Owen's book on incremental OO refactoring.

create Don't test randomness — understand, extract, control it takeaway

kind (empty) → takeaway

name (empty) → Don't test randomness — understand, extract, control it

slug (empty) → don-t-test-randomness-understand-extract-control-it

attrs (empty) → {"type" => "recommendation"}

description (empty) → Rodzik's closing prescription for testing non-deterministic code: don't try to test randomness directly. Instead unde...

short_description (empty) → Closing message: rather than testing random output, isolate randomness behind an abstraction you can control.

create Avoid logic in tests takeaway

kind (empty) → takeaway

name (empty) → Avoid logic in tests

slug (empty) → avoid-logic-in-tests

attrs (empty) → {"type" => "warning"}

description (empty) → Logic inside tests — if/else branches mirroring production, loops that mechanically multiply assertions, or calls to ...

short_description (empty) → Conditionals, loops, and production helpers inside tests produce green-but-wrong suites.

create Inject test doubles via dependency injection takeaway

kind (empty) → takeaway

name (empty) → Inject test doubles via dependency injection

slug (empty) → inject-test-doubles-via-dependency-injection

attrs (empty) → {"type" => "recommendation"}

description (empty) → When stubbing methods couples tests to implementation, extract an interface (e.g. `Dice`) and dependency-inject an im...

short_description (empty) → Replace ad-hoc stubs with an injected fake collaborator behind an extracted interface.

create Log state changes to make tests assert on behavior takeaway

kind (empty) → takeaway

name (empty) → Log state changes to make tests assert on behavior

slug (empty) → log-state-changes-to-make-tests-assert-on-behavior

attrs (empty) → {"type" => "recommendation"}

description (empty) → Recording an event on every state change (roll, assignment, winner decision) gives tests something richer than the fi...

short_description (empty) → Append events on every state change so tests and debuggers can see the full execution trace.

create Keep fake implementations out of production code takeaway

kind (empty) → takeaway

name (empty) → Keep fake implementations out of production code

slug (empty) → keep-fake-implementations-out-of-production-code

attrs (empty) → {"type" => "recommendation"}

description (empty) → Rodzik's personal rule: never ship fake implementations alongside production code, even if it means duplicating them ...

short_description (empty) → Test doubles belong in the test suite, not shipped alongside production files.

create Are fake-dice tests enough, or should the real dice be tested too? question

kind (empty) → question

name (empty) → Are fake-dice tests enough, or should the real dice be tested too?

slug (empty) → are-fake-dice-tests-enough-or-should-the-real-dice-be-tested-too

attrs (empty) → {"answer_summary" => "Unit tests with fake dice are fine for the unit level, but real-world code should also have end...

description (empty) → Audience asks whether tests using the fake dice are sufficient. Rodzik agrees that in the real world you should compl...

short_description (empty) → Q&A on whether unit tests with fake dice need complementary end-to-end tests.

create Is time travel the only way to test Time.current? question

kind (empty) → question

name (empty) → Is time travel the only way to test Time.current?

slug (empty) → is-time-travel-the-only-way-to-test-time-current

attrs (empty) → {"answer_summary" => "No. Algebraic effects and dependency injection of a per-request current-time object are better ...

description (empty) → An audience member challenges Rodzik's assertion that time-travel helpers are the only way to test `Time.current`/`Ti...

short_description (empty) → Challenge to Rodzik's claim that time-travel helpers are the only way to test current-time code.

create Should fake implementations live with production code? question

kind (empty) → question

name (empty) → Should fake implementations live with production code?

slug (empty) → should-fake-implementations-live-with-production-code

attrs (empty) → {"answer_summary" => "Rodzik keeps fakes out of production code even at the cost of duplication. Audience adds: run t...

description (empty) → Audience asks whether reusable test-only objects (fake dice, fake current-time) should live with production files so ...

short_description (empty) → Q&A on where to keep reusable test-only objects like a fake dice or fake time.

Edges (43)

update Jakub Rodzik → authored → Testing Randomness

context (empty) → Presented the talk at wroclove.rb 2023.

update Testing Randomness → presented_at → wroclove.rb 2023

context (empty) → Single-track Ruby conference where the talk was delivered.

create Jakub Rodzik → attended → wroclove.rb 2023

context (empty) → Spoke at the conference.

relation (empty) → attended

source_node_id (empty) → 7d93f41a-54c6-470a-90e3-895b8f2b5f92

target_node_id (empty) → f1adcf6d-780c-4418-8a81-c88da8b4e631

create Testing Randomness → about → Test Driven Development

context (empty) → Fredo begins by writing RSpec tests first and labels his approach TDD.

relation (empty) → about